A few years ago, Nancy Birdsall and I proposed that donors might consider a scheme to give aid to developing countries based on the services they actually deliver. For example, donors could promise to pay $100 for each additional child who completes primary school and takes a standardized competency test. The Center for Global Development has worked further on this idea, and rebranded it as “Cash on Delivery aid“. CGD will soon be publishing a book setting out how the idea might work in practice.
The Conservative Party have said in their Green Paper on international development that they will pursue this approach if they form the next government in the UK. They say:
We will link aid directly to independently-audited evidence of real progress on the ground. Increasingly, we will pay ‘cash on delivery’: giving an agreed amount to a recipient government for every extra child they get into school or every extra person who receives decent healthcare. This will give British taxpayers confidence that their aid money is buying specific successful outcomes.
In October, CAFOD published a briefing note about Cash on Delivery aid which is a helpful summary of the proposal, and is also a useful compilation of some questions that have been raised about the idea. It lists nine “risks” of Cash on Delivery aid which it says should be addressed.
I have written a response to the CAFOD brief which addresses each of the nine risks in turn. I regard nearly all of the issues raised by CAFOD as features, rather than risks, of Cash on Delivery aid.
7 responses to “Cash on Delivery Aid: Response to CAFOD questions”
[…] Owen Barder responds to some good questions from my former NGO, CAFOD about the proposal for making aid Cash on Delivery […]
[…] are very good questions, which Owen Barder tries to address them here. He also makes a stinging assessment of the issues raised by CAFOD, arguing that what the […]
[…] Barder, who with Nancy Birdsall has done much of the work in developing the CoD concept, responded to these concerns on his blog. All I would add is that a) many of the concerns apply equally to other forms of aid […]
COD doesn’t need to work at all times in all places for it to be useful. And it doesn’t need to be free from flaws for it to be worth a shot. Owen, in arguing against CAFOD’s criticisms, goes overboard in arguing the perfection of COD, assuming all best case scenarios that in practice are unlikely to occur. In refusing to take a balanced stand and discuss the pros and cons of COD, he makes those already skeptical of COD even more so.
For example, take the point on aid orphans and darlings. Aid orphans by CAFOD’s definition are clearly not countries or communities that could use more aid productively but can’t get it; rather, they are countries that wouldn’t use aid most effectively, because they don’t have the capacity to do so. Similarly, aid darlings use aid very effectively, much more productively than aid orphans – that’s why people are so eager to give them money. COD would help very little those who have great difficulty improving results, while aid darlings would do very well with COD. CAFOD is right on the mark with the comment that COD would widen the gap between aid orphans and aid darlings if implemented on a wide scale and, as is currently envisioned (at least by my research), with a single, uncontextualised indicator of progress for all countries.
To further illustrate the point: take two hypothetical communities in a single city in my country, Canada. One is poor, with a low level of education, while the other is rich, with a high level of education. The city government is instituting a nutrition and exercise education program to combat obesity and related health problems, which are high in both communities. Say, to spur this program, the national government used a COD approach: for every person who is of a normal weight above a baseline, the city government will receive $1 000. Investment flows to the high income community, as greater levels of education virtually guarantee a more receptive audience to the education program and increased effectiveness. The number of normal weight people in the high income community grows; money flows to the city government; more program investment occurs in the high income community to fuel more positive results. Perhaps at some point the marginal cost of producing one normal weight person in the high income community increases beyond the marginal cost in the low income community, but if there are a very large number of overweight people in the communities, this may not happen for some time. Only in specific circumstances (better investment return, an unbiased, universally concerned government, rationally compassionate policy making, etc.) would it make sense to invest in the low income community, as Owen claims would happen as more COD money flowed into the government.
Just because it is more difficult to create results in one community than another doesn’t mean it’s not worth working in the first community. COD seems to see development outcomes as the same as market outcomes, in which the more money that is earned the better the outcome (the development counterpart is the more people helped and the more significantly those people are helped, in terms of increases in outcomes, the better). That is a highly contestable idea in development and economics, as improving the lives of the worst off or those in the most unequal situation may be of more value than more significantly improving the lives of more people if the later group is wealthier than the former. The fact that COD in its current incarnation favours aid darlings over aid orphans (whether at a country, community or individual level) is a real and likely consequence, and Owen should treat it more seriously.
That doesn’t mean COD should be abandoned. No single aid ‘modality’ (to use the lingo) will fix all problems. COD doesn’t need to be perfect to be useful. If we acknowledge the weaknesses as well as the strengths of the idea and provide concrete arguments about where it should be used and where it shouldn’t, COD is much more likely to secure widespread support.
@Duncan – I’m sorry you think that I am arguing that COD is, or would be, perfect; or that I refuse to discuss the advantages and disadvantages. That certainly isn’t my intention. I think the idea is worth testing, but I would never claim that it is the answer to everything, or that it is without weaknesses.
You make a good point about the risk that COD aid (or any results-based financing mechanism) will encourage governments to invest most in the easiest to reach people or communities. (I think of this as the ‘low hanging fruit’ problem rather than the ‘aid orphans and darlings’ problem, which I understand as a slightly different problem to do with misallocation of donor resources across countries; but whatever we call it, it is a valid issue to raise.)
There are some circumstances in which it is desirable to scale up most rapidly to provide services where they are easier or cheaper. For example, in a country that is moving to universal primary education, the community there may conclude that there is no point in insisting that kids living in periurban areas should have to wait for their education until other children, more remote and harder to reach, are also being served. But as you rightly say, there are times when we are concerned to help one community in particular, perhaps to try to redress a problem of social exclusion or marginalisation, or to remedy a historical injustice. It is perfectly reasonable for a society to place more value on helping someone who is relatively poor than helping someone who is relatively rich. Social justice is an entirely valid policy objective.
We achieve this in a framework of results based aid such as Cash on Delivery by adjusting the payments to reflect these social values. If we believe that getting girls into school is more valuable than getting boys into school, we can have a higher payment for girls than boys. If we believe that getting clean drinking water to a marginalised community is more important than getting clean water in the main city, we can offer to pay more (based on region, or distance from the city).
In your example of an obesity programme, a government might decide either (a) they want to reduce obesity by as many people as possible, irrespective of who they are, in which case a uniform COD approach might be appropriate; or (b) they also care about the distribution of obesity, or of the targeting of the programme, so they offer a larger payment for participation in the low income community. Both approaches seem to me to have something to commend them – they are a matter of a community’s values. The job of economics in this case is to observe that there is a trade-off – for a given budget, by targeting the low income community the result is that fewer people are reached overall. I would only be uncomfortable if the decision were taken without recognising that this trade-off is real, or if the reduction in the total needed to target the programme on low income households becomes so great that the society no longer believes that it is worthwhile.
An advantage of achieving this goal by setting different rewards in a Cash on Delivery programme is that it would this trade-off explicit. As long as it is transparent, this ensures that the choice you describe is made consciously, and it makes it more likely that the choice really does represent society’s values and not merely the preferences of vested interests or elites.
Owen
Thanks for responding to my post, Owen. I am glad you mentioned the tradeoff in CoD (and in aid and development more generally) between helping those that are easiest to help and those that are in the worst circumstances, as only in the rarest circumstances are those groups the same. In my post I simply wanted to point out that they tradeoff is as real in CoD as in any other form of aid. Your potential solutions for the low-hanging fruit problem using CoD are intriguing, and so long as they did not detract too much from the attractive simplicity of CoD and the transparency and reduced transactions costs that come with that simplicity, the solutions you mention seem like they could be effective.
Perhaps at some point, if CoD is proven to work, donors will leave everything to the recipients, including defining the priority and the indicator. Then we’ll see the benefits of true recipient country ownership, for goals, design and implementation. But I suppose we have to test CoD first.
[…] main piece of advice on development is to go for “cash on delivery”. Pay a non-inclusive government only for goods and services it has delivered, not for the process. […]